When a production computer dies

BitDepth#1348

MARK LYNDERSAY

AS BUSINESSES become increasingly digital, the architecture of the computing systems that underpin them becomes increasingly important.

So what happens when a key part of the network fails?

I didn't plan for what happened on March 26, but there were enough paranoia-inspired buffers in place to compensate for the failure of the core of my IT infrastructure, the imaging workstation around which much of my day-to-day work revolves.

The failed system is a Mac Pro tower, which is connected to everything I work with and hosted five hard drives and two SSDs representing a primary resource of 50 TB.

The drives are fine and I can access them on an emergency basis. A system of drive bays allows me to mount the bare drives and retrieve critical files. But that workflow ecosystem is not available until replacement equipment arrives.

Here's how I addressed the disruption.

Evaluate

What is lost? What is salvageable? Because I don't use complicated drive schemes like RAID redundancy, the system failure leaves me with disks that must be remounted and their digital relationship re-established. The new architecture connects the disks the same way, just using different hardware to do it. OS level permission issues shouldn't arise.

How to restore function?

The easiest thing to do would have been to replace the tower. They remain in demand and are pricey for their age, but not prohibitively so.

But along with my experience, a colleague who also prizes the towers for film editing has experienced terminal hardware failure in the last two years.

My tower was manufactured in 2009. The last of that highly-regarded second-generation Mac Pro was made in 2013.

Replacement parameters

Many IT pros have stories to tell about parts of their network that can't be touched. Software that can't run on modern operating systems. Hardware that must be connected to specific peripherals using ports that don't exist any more.

For now, the system needs to remain on Mac OS High Sierra (2017), so any new box would have to be able to run that OS.

It turns out most of the Mac hardware made after 2016 won't run it.

How will the new architecture address the old problem?

Any major IT infrastructure change must either maintain a working status quo or improve it.

The Mac Mini of that vintage is affordable and agreeably small, but don't have the video horsepower to drive my monitors. The third-generation Mac Pro, the so-called trashcan Mac, remains expensive and is expensive to upgrade.

An agreeable configuration of a late 2015 iMac came very close to the computing engine I'd been working with at a price, after upgrades, that was equal to the starting point for the trashcan Mac.e.

SWOT analysis

Strengths. I expect roughly equivalent performance with dramatically lower power consumption, even with external drive boxes.

Weaknesses. Making a working drive architecture external makes it more modular, but also slower and more prone to accidental disconnection.

Opportunities. A chance to rethink my workspace, because that Mac Pro was big. The migration from Adobe's products that I've charted and the use of tools that demand the older OS must be accelerated.

Threats. The new computer isn't ancient, but can't be described as new. The Mac Pro ran 24/7 for ten years. I won't expect that from an iMac. My window for this new configuration is five years.

What I learned

Abstracting data from hardware was an intuitive decision for this system configuration, but modularity is something to more intentionally build into future versions of the workstation and its satellite connections and dependencies.

Mark Lyndersay is the editor of technewstt.com. An expanded version of this column can be found there